Eyesight is one of our most entrusted senses, constituting a key element of most of our daily activities. Nonetheless, the reliability of our visual perception can be challenged by a psychological phenomenon called “change blindness”, which refers to our inability to notice changes in a scene, even when they are seemingly obvious (Hollingworth, 2006). This phenomenon has received great interest by researchers, as it presents an opportunity to examine the conditions that need to be satisfied for percepts to access our conscious awareness (Noe, 2005).
A common method of demonstrating and studying change blindness is the flicker paradigm (Rensink et al., 1997). In this paradigm, participants are shown two images of a scene (one original and one modified) that switch back and forth, separated by a short, blank screen (approximately 100 milliseconds). Participants are asked to report the change in the scene as soon as they notice it. If you would like to try it for yourself, you may find two examples here: Video 1 and Video 2.
As you can see in Video 2, widely known as the “airplane” example (Rensink et al., 1997), these changes can be as apparent and crucial as the removal of an airplane’s engine, although they require a significant amount of time to be noticed. Indeed, the series of experiments by Rensink and colleagues (1997), which tested different versions of the flicker paradigm, has shown that hardly any changes were noticed during the initial round of alternation, while most changes remained undetected after almost one minute of viewing.
The Mechanism Behind the Phenomenon
Several mechanisms have been proposed to illuminate change blindness, ranging from age-related differences (Bergmann et al., 2016) to reflexive attentional shifts (Smith & Schenk, 2008). While multiple factors likely contribute to this phenomenon, two explanations in particular have received much support. First mechanism refers to “visual transients”. That is, the blank screen that flickers between the two images leads to very rapid changes in luminance and colour, which are immediately detected by the retina. Due to the automatic attention attracted by this motion, other changes in the scene require longer to be noticed. Supporting this explanation, in trials where the flickering blank screen was eliminated, participants were significantly faster at recognising the modification, since the disruption of the global transient was absent (Rensink et al., 1997; Tarr & Aginsky, 1996).
A further implication of this mechanism is that other types of global transients should induce a similar effect of change blindness. This can be tested via O’Regan and colleagues’ (1999) mudsplashes paradigm, in which a small number of occluding shapes are randomly placed on the two alternating images. Addressing a limitation of the flicker paradigm, this method does not occlude the location of the modified part of the image, hence making sure that the modification is visible and can be registered by the visual system. You may find an example of the mudsplashes paradigm through this video.
Similar to the flicker paradigm, O’Regan et al. (1999) revealed that viewers were still virtually blind to the change in the image with the mudsplashes paradigm. This indicates that if a brief visual disturbance is introduced whilst a change is occurring, this disturbance distracts away from the change, causing it to be more difficult to detect (Turatto et al., 2003).
Alongside the visual transient explanation, a simultaneous mechanism at play is related to our limited visual working memory (Brady et al., 2011). In order to notice a change, we need to be able to encode the original image into our memory and compare it with the modified version. Nonetheless, the very rapid display times of the images, as well as our brains’ limited capacity, hinder the encoding of every detail into our memory, causing some of the visual information to be discarded after being processed (Escamilla et al., 2020).
Criticisms and Recent Advancements
Although change blindness has been studied for decades, an important criticism is that the phenomenon has only been tested qualitatively and via 2D images, which differ drastically from 3D scenes. For example, compared to 2D images, immersive 3D environments enable testing attention and memory under more naturalistic viewing conditions, making findings more generalisable to our daily visual experience (Hayhoe & Rothkopf, 2011; Li et al., 2016). In addition to its greater external validity, spatial learning tasks in 3D allow observers to also integrate proprioceptive and vestibular feedback, while feedback from their full-body motion helps strengthen the relationship between the objects and the observer (Chrastil & Warren, 2012; Li et al., 2016). Most of the literature neglects these crucial components of memory and attention due to their heavy reliance on 2D stimuli.
With the introduction of virtual reality (VR) technologies into psychological research, recent studies have been able to investigate this phenomenon in 3D tasks. One such study has been conducted by Martin and colleagues (2023), which investigated a number of change manipulations (e.g., distance, type of manipulation, field of view, complexity of the modified element) in an immersive 3D VR environment. Participants were asked to observe a realistic, stereoscopic virtual scene of a living room with several furniture. In each trial, one element in the scene was manipulated, with modifications ranging from addition (a new object is added to the scene) to relocation (a pre-existing object is moved to a different location), alongside the aforementioned change properties. To re-create the effect of a flicker, the changed element switched between its original and manipulated states every time the observer rotated their head and the element remained outside of their field of view.
Results revealed a detection ratio of 41.97%, meaning that even though participants were explicitly asked to identify a change for 45 seconds, more than half of the manipulations went undetected. This constitutes a much higher ratio of unnoticed changes compared to past research with 2D stimuli, which have found detection ratios ranging between 73-92% (Hollingworth, 2006; 2007; Vasser et al., 2015). Furthermore, unlike the results obtained with 2D images, Martin et al. (2023) revealed that the type of the manipulation did not significantly affect its detectability.
These findings suggest that the change blindness effect might function differently in 3D immersive environments that more closely resemble our real-life observing conditions. Preliminary findings with VR indicate that limitations in our visual memory continue to influence our detection of changes, even more so than 2D modifications. Future work with 3D content is integral to gain a better understanding of the types of manipulations that remain undetected by human observers. The knowledge gained by researching change blindness in VR can have important applications for several areas, including redirected walking (Sun et al., 2018), interior design (Alghofaili et al., 2019), cinematography (Serrano et al., 2017), and gaze prediction (Martin et al., 2022).
References
Alghofaili, R., Solah, M. S., Huang, H., Sawahata, Y., Pomplun, M., & Yu, L. (2019). Optimizing visual element placement via visual attention analysis. 2020 IEEE Conference on Virtual Reality and 3D User Interfaces (VR), 464–473. https://doi.org/10.1109/vr.2019.8797816
Bergmann, K., Schubert, A., Hagemann, D., & Schankin, A. (2016). Age-related differences in the P3 amplitude in change blindness. Psychological Research, 80(4), 660–676. https://doi.org/10.1007/s00426-015-0669-6
Brady, T. F., Konkle, T., & Alvarez, G. A. (2011). A review of visual memory capacity: Beyond individual items and toward structured representations. Journal of vision, 11(5), 4-4.
Chrastil, E. R., & Warren, W. H. (2014). Active and passive spatial learning in human navigation: Acquisition of graph knowledge. Journal of Experimental Psychology Learning Memory and Cognition, 41(4), 1162–1178. https://doi.org/10.1037/xlm0000082
Escamilla, J. C., Castro, J. J. F., Baliyan, S., Ortells-Pareja, J. J., Rodríguez, J. J. O., & Cimadevilla, J. M. (2020). Allocentric Spatial Memory Performance in a Virtual Reality-Based Task is Conditioned by Visuospatial Working Memory Capacity. Brain Sciences, 10(8), 552. https://doi.org/10.3390/brainsci10080552
Hayhoe, M. M., & Rothkopf, C. A. (2011). Vision in the natural world. Wiley Interdisciplinary Reviews Cognitive Science, 2(2), 158–166. https://doi.org/10.1002/wcs.113
Hollingworth, A. (2006). Scene and position specificity in visual memory for objects. Journal of Experimental Psychology Learning Memory and Cognition, 32(1), 58–69. https://doi.org/10.1037/0278-7393.32.1.58
Hollingworth, A. (2007). Object-position binding in visual memory for natural scenes and object arrays. Journal of Experimental Psychology Human Perception & Performance, 33(1), 31–47. https://doi.org/10.1037/0096-1523.33.1.31
Li, C., Aivar, M. P., Kit, D. M., Tong, M. H., & Hayhoe, M. M. (2016). Memory and visual search in naturalistic 2D and 3D environments. Journal of Vision, 16(8), 9. https://doi.org/10.1167/16.8.9
Martin, D., Serrano, A., Bergman, A. W., Wetzstein, G., & Masia, B. (2022). Scangan360: A generative model of realistic scanpaths for 360 images. IEEE Transactions on Visualization and Computer Graphics, 28(5), 2003-2013.
Martin, D., Sun, X., Gutierrez, D., & Masia, B. (2023). A study of Change blindness in immersive environments. IEEE Transactions on Visualization and Computer Graphics, 29(5), 2446–2455. https://doi.org/10.1109/tvcg.2023.3247102
Noe, A. (2005). What does change blindness teach us about consciousness?. Trends in Cognitive Sciences, 9(5), 218-218.
O’Regan, J. K., Rensink, R. A., & Clark, J. J. (1999). Change-blindness as a result of ‘mudsplashes’. Nature, 398(6722), 34-34. https://doi.org/10.1038/17953
Rensink, R. A., O’Regan, J. K., & Clark, J. J. (1997). To See or not to See: The Need for Attention to Perceive Changes in Scenes. Psychological Science, 8(5), 368–373. https://doi.org/10.1111/j.1467-9280.1997.tb00427.x
Serrano, A., Sitzmann, V., Ruiz-Borau, J., Wetzstein, G., Gutierrez, D., & Masia, B. (2017). Movie editing and cognitive event segmentation in virtual reality video. ACM Transactions on Graphics, 36(4), 1–12. https://doi.org/10.1145/3072959.3073668
Smith, D. T., & Schenk, T. (2008). Reflexive attention attenuates change blindness (but only briefly). Perception & Psychophysics, 70(3), 489–495. https://doi.org/10.3758/pp.70.3.489
Sun, Q., Patney, A., Wei, L., Shapira, O., Lu, J., Asente, P., Zhu, S., Mcguire, M., Luebke, D., & Kaufman, A. (2018). Towards virtual reality infinite walking. ACM Transactions on Graphics, 37(4), 1–13. https://doi.org/10.1145/3197517.3201294
Tarr, M., & Aginsky, V. (1996). From objects to scenes: Speculations on similarities and differences. In Scene Recognition Workshop, Max-Planck-Institut für Biologische Kybernetik, Tübingen.
Turatto, M., Bettella, S., Umiltà, C., & Bridgeman, B. (2003). Perceptual conditions necessary to induce change blindness. Visual Cognition, 10(2), 233–255. https://doi.org/10.1080/713756677
Vasser, M., Kängsepp, M., & Aru, J. (2015). Change blindness in 3d virtual reality. arXiv preprint arXiv:1508.05782.
